| dc.description.abstract | Recently development methodologies have been proposed which describe themselves as "Object Oriented". While all of them offer approaches to extended data and behavioural modelling, none of them seem fully adequate to address the total concept of object-oriented development. They often do not provide constructs which lead to the use of databases, nor do they always recognise the shift from sequential to prototyping style which is inherent in much object-oriented technology. The objective of this thesis is to lay a framework for an object-oriented methodology suitable for OODBMS. Details of conventional methods for developing database applications, and of the recent OO methods, have been examined and compared in order to propose a coherent set of tasks and deliverables. Account has also been taken of designing for re-use, which has been one of the main selling points of the OO approach. The proposed methodology attempts to address related side issues, with particular focus on object concurrency, which seems particularly thinly covered in many of the current proposals. Many other side issues are also mentioned, but due to time constraints, they are not given any further discussion. The topic is an extremely multidisciplinary one, and a very wide range of expertise would be necessary to do justice to all these aspects. Mapping of the new methodology has been tried on two case study examples using Postgres and Ontos. Postgres is an extended relational DBMS developed as a research prototype at University of California, Berkeley. Ontos is the commercial object-oriented DBMS marketed by Ontos Incorporated, Burlington, Massachusetts. Some details of these implementation examples are included. | en_US |
